cognitive-load
Improves technical communication by breaking complex information into digestible chunks to respect human working memory limits.
Install
mkdir -p .claude/skills/cognitive-load && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/15941" && unzip -o skill.zip -d .claude/skills/cognitive-load && rm skill.zipInstalls to .claude/skills/cognitive-load
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Don't overwhelm — chunk, scaffold, summarize first.Key capabilities
- →Chunk complex responses into 3-5 logical sections
- →Minimize extraneous cognitive load by reducing jargon and irrelevant details
- →Maximize germane cognitive load by using analogies and examples
- →Implement progressive disclosure for information delivery
- →Start complex explanations with a summary
- →Use guard clauses and flat structures in code examples to reduce load
How it works
The skill applies Miller's Law by structuring responses into digestible chunks, summarizing first, and progressively disclosing details, while minimizing extraneous information and maximizing germane load through examples.
Inputs & outputs
When to use cognitive-load
- →Simplify complex architectural explanations into manageable sections
- →Structure long-form technical documentation for better readability
- →Explain unfamiliar programming concepts using analogies and scaffolding
- →Reduce jargon and extraneous detail in code walkthroughs
About this skill
Cognitive Load Skill
Don't overwhelm — chunk, scaffold, summarize first.
Core Principle
Working memory holds 4±1 items (Miller's Law). Exceed this → comprehension drops, frustration rises. Every explanation must respect this limit.
Cognitive Load Types
| Type | Definition | Goal | Example |
|---|---|---|---|
| Intrinsic | Inherent task complexity | Manage via scaffolding | Learning recursion is inherently complex |
| Extraneous | Load from poor presentation | Minimize aggressively | Cluttered UI, jargon, irrelevant details |
| Germane | Load from building mental models | Maximize | Analogies, examples, connections to prior knowledge |
Key insight: We can't reduce intrinsic load, but we can minimize extraneous and maximize germane.
Chunking Strategies
The 3-5 Rule
Break any complex response into 3-5 logical chunks. Each chunk should be digestible in isolation.
## Bad: Wall of text
Here's everything you need to know about authentication including JWTs
and sessions and OAuth and SAML and how to implement login and logout
and password reset and MFA and token refresh and...
## Good: Chunked
### 1. Authentication Basics
Brief explanation of what authentication is.
### 2. Token-Based (JWT)
Just the JWT pattern.
### 3. Session-Based
Just the session pattern.
### 4. When to Use Which
Comparison table.
Group Related Items
Present related concepts together, separated from unrelated ones.
## Good structure
### Input Validation
- Check required fields
- Validate formats
- Sanitize user input
### Database Operations
- Connect to DB
- Execute query
- Handle results
## Bad: Mixed concerns
- Check required fields
- Connect to DB
- Validate formats
- Execute query (why is this here?)
Progressive Disclosure
Level 1: Summary (everyone gets this) Level 2: Details (for those who want more) Level 3: Implementation (only when requested)
## Level 1: Summary
OAuth is a protocol that lets users grant third-party apps access
to their data without sharing passwords.
## Level 2: Details (expand if asked)
OAuth 2.0 defines four roles: Resource Owner, Client, Authorization
Server, and Resource Server. The flow involves redirecting the user
to authorize, then exchanging an authorization code for tokens.
## Level 3: Implementation (only if specifically requested)
[Full code example with error handling, token storage, refresh logic]
Always ask before going deeper: "Want me to dive into the implementation details?"
Summarize First Pattern
Start EVERY complex explanation with a summary. Then optionally expand.
## Summary
Authentication verifies who you are. Authorization determines what you
can access. JWT is a stateless token format. Sessions are server-stored.
## Details
[Only if user wants more]
High Load vs Low Load Presentation
| High Cognitive Load | Low Cognitive Load |
|---|---|
| Wall of text | Headers + bullets |
| Multiple concepts at once | One concept at a time |
| Technical jargon | Plain language first, then terms |
| Deep nesting (if > if > if) | Early returns, flat structure |
| Abstract first | Concrete example first |
| Long code blocks | Focused snippets |
| No visual breaks | Whitespace + visual hierarchy |
Code Example: High vs Low Load
// HIGH LOAD: Too much at once
function processUserRegistration(data) {
if (data.email && data.email.includes('@') && data.email.length > 5) {
if (data.password && data.password.length >= 8 && /[A-Z]/.test(data.password) && /[0-9]/.test(data.password)) {
if (data.age && data.age >= 18) {
// 3 levels deep, reader lost context
}
}
}
}
// LOW LOAD: Guard clauses, flat structure
function processUserRegistration(data) {
if (!isValidEmail(data.email)) {
return { error: 'Invalid email' };
}
if (!isStrongPassword(data.password)) {
return { error: 'Password too weak' };
}
if (!isAdult(data.age)) {
return { error: 'Must be 18+' };
}
return createUser(data);
}
Overload Signals and Responses
| User Signal | What It Means | Your Response |
|---|---|---|
| "I'm confused" | Extraneous load too high | Stop, simplify, use analogy |
| Repeated questions | Intrinsic load not scaffolded | Step back to fundamentals |
| Short/frustrated responses | Overwhelmed | Acknowledge, offer break or simpler path |
| "Just tell me how" | Wants action, not theory | Skip explanation, give steps |
| Silent/no follow-up | May be processing OR lost | Check in: "Does that help?" |
3-3-3 Rule
A quick heuristic for any explanation:
- 3 sentences for the summary
- 3 examples to make it concrete
- 3 minutes before checking understanding
If the user hasn't responded after a complex explanation, pause and check in rather than continuing to add load.
Scaffolding Complex Topics
For high-intrinsic-load topics, build up gradually:
### Step 1: Analogy to familiar concept
"A database index is like a book's index — it lets you find
content without reading every page."
### Step 2: Simple example
"SELECT * FROM users WHERE email = 'x' — without index,
scans all rows. With index, jumps directly."
### Step 3: The real complexity
"B-tree indexes balance lookup speed vs. write overhead..."
Self-Check Before Responding
Before sending a complex response, ask:
- Did I summarize first?
- Is it chunked into 3-5 sections?
- Am I introducing only ONE new concept per section?
- Can I replace jargon with plain language?
- Would an example make this clearer?
If any answer is "no", revise before sending.
When not to use it
- →When the user explicitly requests all details at once
- →When the information is inherently simple and does not require chunking
- →When the user is already familiar with the topic and needs direct answers
Limitations
- →Working memory holds 4±1 items (Miller's Law)
- →Cannot reduce intrinsic task complexity
- →Requires user interaction to determine when to dive deeper into details
How it compares
This workflow systematically structures AI responses to align with human working memory limits, actively reducing cognitive overload through chunking and progressive disclosure, unlike a generic response that might present all information a
Compared to similar skills
cognitive-load side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| cognitive-load (this skill) | 0 | 3mo | No flags | Advanced |
| user-file-ops | 2 | 3mo | Review | Beginner |
| history-insight | 0 | 7mo | Review | Beginner |
| meeting-minutes | 41 | 6mo | No flags | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by fabioc-aloha
View all by fabioc-aloha →You might also like
user-file-ops
trpc-group
Simple operations on user-provided text files including summarization.
history-insight
team-attention
This skill should be used when user wants to access, capture, or reference Claude Code session history. Trigger when user says "capture session", "save session history", or references past/current conversation as a source - whether for saving, extracting, summarizing, or reviewing. This includes any mention of "what we discussed", "today's work", "session history", or when user treats the conversation itself as source material (e.g., "from our conversation").
meeting-minutes
github
Generate concise, actionable meeting minutes for internal meetings. Includes metadata, attendees, agenda, decisions, action items (owner + due date), and follow-up steps.
knowledge-absorber
YuJunZhiXue
深度解析链接、文档或代码,生成“全能导师级”的教学笔记(零基础直达精通)。具备“真理锚定”校验能力,自动识别幻觉与过时信息。
academic-paper-review
q2805187159
Use this skill when the user requests to review, analyze, critique, or summarize academic papers, research articles, preprints, or scientific publications. Supports comprehensive structured reviews covering methodology assessment, contribution evaluation, literature positioning, and constructive fee
docx
anthropics
Comprehensive document creation, editing, and analysis with support for tracked changes, comments, formatting preservation, and text extraction. When Claude needs to work with professional documents (.docx files) for: (1) Creating new documents, (2) Modifying or editing content, (3) Working with tracked changes, (4) Adding comments, or any other document tasks